1. Technical Field
The present disclosure relates to a parameter determining apparatus for estimating temperature of a switching element of an inverter.
2. Description of the Related Art
Generally, a multi-purpose inverter is configured with a rectification unit, a direct current (DC) link capacitor, and an inverter unit, and the inverter unit is configured with a power semiconductor module. Such a power semiconductor module turns on and off a current while an inverter operates such that heat is generated.
When temperature of the power semiconductor module rises over a predetermined temperature, a switching element configuring the power semiconductor module is damaged so that a temperature rising is a major consideration factor in selecting the power semiconductor module.
FIG. 1 is a cross sectional view for describing a configuration of a conventional power semiconductor module.
A power semiconductor module 200 is arranged on a printed circuit board (PCB) 100, and a heatsink 300 is arranged below the PCB 100 to radiate heat generated from the power semiconductor module 200. A plurality of switching elements 220 of a chip shape are arranged on an inner substrate 210 inside the power semiconductor module 200. Each of the switching elements 220 is, for example, an insulated gate bipolar transistor (IGBT). In such a configured power semiconductor module 200, a heating point of highest temperature is a junction A inside the switching element 220, but the junction A is located inside the power semiconductor module 200 so that temperature could not be directly measured. Therefore, the temperature of the junction A of the switching element 220 is indirectly estimated and a margin is added to the estimated temperature, thereby preventing a temperature rising of the junction A.
A method for indirectly estimating temperature of the junction A is to calculate a loss based on temperature of a position (for example, B of FIG. 1) at which a temperature measurement may be possible. Then, a temperature difference is calculated using the calculated loss and a thermal resistance between the position at which the temperature is measured and the junction of the switching element 220, thereby estimating temperature of a junction. When the loss is calculated during such a process, an intrinsic parameter of the switching element 220 is used and this intrinsic parameter is experimentally provided by each of manufacturers. That is, the loss is estimated using a parameter under a specific condition among parameters varied according to a temperature variation.
To estimate temperature of a junction of a switching element, a process of calculating a loss of the switching element 220 is as follows. It will be described a case in which the switching element 220 is an IGBT.
In other words, a temperature difference is calculated by adding a conduction loss and a switching loss of the switching element 220 to a conduction loss and a switching loss of an anti-parallel diode of the IGBT and by multiplying the added result by thermal resistance provided from a manufacturer, and then a temperature difference from a measured reference temperature is added to the calculated temperature difference to estimate temperature of a junction of the IGBT. At this point, the following Equations are used.
                              P          C_IGBT                =                                            V                              CE                ⁢                                                                  ⁢                0                                      ⁢                                          I                0                            ⁡                              (                                                      1                                          2                      ⁢                                                                                          ⁢                      π                                                        +                                                                                    m                        a                                            ⁢                      cos                      ⁢                                                                                          ⁢                                              ϕ                        1                                                              8                                                  )                                              +                                    R              CE                        ⁢                                          I                0                2                            ⁡                              (                                                      1                    8                                    +                                                                                    m                        a                                            ⁢                      cos                      ⁢                                                                                          ⁢                                              ϕ                        1                                                                                    3                      ⁢                                                                                          ⁢                      π                                                                      )                                                                        [                  Equation          ⁢                                          ⁢          1                ]                                          P                      sw_IGB            ⁢                                                  ⁢            T                          =                                            (                                                                    E                    on                                    ⁡                                      [                                                                  I                        0                                            π                                        ]                                                  +                                                      E                    off                                    ⁡                                      [                                                                  I                        0                                            π                                        ]                                                              )                        ·                          f              sw                                =                                                                 [                                                                            (                                                                        A                          +                          L                                                                          π                          2                                                                    )                                        ⁢                                          I                      0                      2                                                        +                                                            (                                                                        B                          +                          M                                                π                                            )                                        ⁢                                          I                      0                                                        +                                      (                                          C                      +                      N                                        )                                                  ]                            ·                              f                sw                                                                        [                  Equation          ⁢                                          ⁢          2                ]            
In Equations 1 and 2, VCEO, RCE, Eon, Eoff are parameters of the power semiconductor module 200, which are provided from a manufacturer. At this point, an initial collector-emitter voltage VCEO and an initial collector-emitter resistance RCE, which are parameters used in Equation 1 for calculating a conduction loss PC_IGBT, may be determined from a graph that is provided from the manufacturer as shown in FIG. 2.
FIG. 2 is a characteristic graph of an IGBT according to temperature, which is provided from a manufacturer, and FIG. 3 is an exemplary diagram for describing a process of determining a parameter in FIG. 2.
Because a parameter used in calculation of a conduction loss is nonlinear, a value corresponding to a condition is obtained from a graph shown in FIG. 2, or a slope value (resistance) and an x intercept (VCEO) configuring a first order function are obtained through a linearization as shown in FIG. 3, thereby obtaining the parameter.
As shown in FIG. 2, the parameter used in the loss calculation is varied according to temperature. Therefore, the manufacturer provides a parameter based on three temperatures (for example, 25° C., 125° C., and 150° C.). When the parameter is directly obtained from a graph, an appropriate point may be determined according to temperature, whereas, when the parameter is obtained through a processor, all information with respect to the parameter should be provided so as to obtain an appropriate value.
Also, even when FIG. 3 is used, parameters required for configuring a graph with respect to an appropriate temperature should be stored in a memory, and an equation linearized through a parameter suitable for a corresponding temperature should also be stored.
Consequently, there are problems in that an amount of data to be stored in a memory is increased so as to consider temperature variation and a calculation process of the data is complicated.